Models / API pricing
LLM API pricing
What every model costs per million tokens, what you get for it, and what a real task costs. Prices are each developer's own list price where we have read it, otherwise the typical host on OpenRouter, with the cheapest host alongside. For how long each model takes, see LLM API speed.
- Models priced
- 214
- Best value
- 13
- no model is both cheaper and more capable
- Prices checked
- 10 Nov 2026
- Blended price
- 3 input tokens to 1 output
Every model's price
Dollars per million tokens. The price is the developer's own list price where we have read it; otherwise it's the typical (median) host on OpenRouter. "Cheapest host" is the cheapest route on OpenRouter for a typical mix, often a compressed (quantised) version of the model. Intelligence and tokens/s are from Artificial Analysis, which doesn't measure every model. Models with a shutdown date announced are left out.
No model matches . Try part of a name or a developer, such as "mistral", or .
| Model | Blended | Input | Output | Cheapest host blended | Intelligence | Our overall | Tokens/s |
|---|---|---|---|---|---|---|---|
| Mistral NemoMistral AI · open weights · typical host price | $0.029 | $0.029 | $0.030 | $0.018Io Net, FP16 · 5 hosts | – | – | – |
| Gemma 3 4BGoogle · open weights · typical host price | $0.042 | $0.040 | $0.050 | $0.042Parasail, BF16 · 2 hosts | 4.8 | – | – |
| gpt-oss-20bOpenAI · open weights · typical host price | $0.058 | $0.030 | $0.14 | $0.022PrimeIntellect, FP4 · 11 hosts | 10.0 | – | 277 |
| Llama 3.1 8B InstructMeta · open weights · typical host price | $0.058 | $0.050 | $0.080 | $0.025DeepInfra, FP8 · 4 hosts | 6.9 | – | – |
| Mistral Small 3Mistral AI · open weights · typical host price | $0.058 | $0.050 | $0.080 | one host | 6.7 | – | – |
| Nova Micro 1.0Amazon · typical host price | $0.061 | $0.035 | $0.14 | $0.061Amazon Bedrock · 2 hosts | 5.9 | – | 260 |
| Command R7B (12-2024)Cohere · typical host price | $0.066 | $0.037 | $0.15 | one host | – | – | – |
| Gemma 3 12BGoogle · open weights · typical host price | $0.075 | $0.050 | $0.15 | $0.075DeepInfra, BF16 · 2 hosts | 3.8 | – | – |
| Granite 4.2 8BIBM · open weights · typical host price | $0.082 | $0.060 | $0.15 | $0.082CoreWeave, BF16 · 2 hosts | 11.1 | – | 65 |
| Phi-4Microsoft · open weights · typical host price | $0.088 | $0.070 | $0.14 | one host | 5.9 | – | 45 |
| Ministral 3 3BMistral AI · open weights · developer price | $0.10 | $0.10 | $0.10 | $0.10Mistral · 2 hosts | 4.8 | – | 209 |
| Nemotron 3 Nano 30B A3BNVIDIA · open weights · typical host price | $0.10 | $0.060 | $0.24 | one host | 8.9 | – | 238 |
| Nova Lite 1.0Amazon · typical host price | $0.10 | $0.060 | $0.24 | $0.10Amazon Bedrock · 2 hosts | 6.7 | – | – |
| Qwen3.5-9BAlibaba · open weights · typical host price | $0.11 | $0.10 | $0.15 | $0.092Darkbloom, FP4 · 6 hosts | 13.3 | – | 110 |
| Qwen3.5-FlashAlibaba · typical host price | $0.11 | $0.065 | $0.26 | one host | – | – | – |
| Llama 3.2 3B InstructMeta · open weights · typical host price | $0.12 | $0.050 | $0.33 | $0.12Parasail, BF16 · 2 hosts | 5.7 | – | – |
| Qwen3-Coder-30B-A3B-InstructAlibaba · open weights · typical host price | $0.12 | $0.070 | $0.28 | $0.12SiliconFlow, FP8 · 2 hosts | 9.6 | – | – |
| Qwen2.5-7B-InstructAlibaba · open weights · typical host price | $0.12 | $0.10 | $0.20 | one host | – | – | – |
| Reka Flash 3Reka AI · open weights · typical host price | $0.12 | $0.10 | $0.20 | one host | 5.6 | – | 83 |
| Qwen3-14BAlibaba · open weights · typical host price | $0.13 | $0.10 | $0.22 | $0.13NextBit, INT4 · 2 hosts | 6.7 | – | – |
| Mistral Small 3.2 24BMistral AI · open weights · typical host price | $0.13 | $0.094 | $0.25 | $0.11DeepInfra, FP8 · 4 hosts | 7.0 | – | – |
| GPT-5 nanoOpenAI · typical host price | $0.14 | $0.050 | $0.40 | $0.14Azure · 3 hosts | 13.0 | – | – |
| Gemma 4 26B A4BGoogle · open weights · typical host price | $0.14 | $0.090 | $0.30 | $0.068Darkbloom · 15 hosts | 16.7 | – | 82 |
| Qwen3-30B-A3B-Instruct-2507Alibaba · open weights · typical host price | $0.14 | $0.090 | $0.30 | $0.14DekaLLM · 3 hosts | 7.5 | – | – |
| GLM-4.7-FlashZ.ai · open weights · typical host price | $0.15 | $0.060 | $0.40 | $0.15Cloudflare · 2 hosts | 14.9 | – | – |
| Gemma 3 27BGoogle · open weights · typical host price | $0.15 | $0.080 | $0.35 | $0.10DeepInfra, FP8 · 4 hosts | 4.9 | – | – |
| Ministral 3 8BMistral AI · open weights · developer price | $0.15 | $0.15 | $0.15 | $0.15Mistral · 2 hosts | 5.5 | – | 93 |
| DeepSeek-V4-Flash (0423)DeepSeek · open weights · typical host price | $0.15 | $0.050 | $0.45 | $0.10StreamLake, FP8 · 13 hosts | 24.4 | – | – |
| Llama 4 ScoutMeta · open weights · typical host price | $0.15 | $0.10 | $0.30 | $0.15DeepInfra, FP8 · 2 hosts | 6.5 | – | 132 |
| Step 3.5 FlashStepFun · open weights · typical host price | $0.15 | $0.10 | $0.30 | one host | 16.6 | – | – |
| Particle-1Particle AI · typical host price | $0.16 | $0.050 | $0.50 | one host | – | 29 | – |
| Nemotron 3 Super 120B A12BNVIDIA · open weights · typical host price | $0.17 | $0.080 | $0.45 | $0.16DeepInfra, BF16 · 3 hosts | 12.8 | – | – |
| GPT-4.1 nanoOpenAI · typical host price | $0.18 | $0.10 | $0.40 | $0.18Azure · 2 hosts | 7.8 | – | – |
| Gemini 2.5 Flash-LiteGoogle · developer price | $0.18 | $0.10 | $0.40 | one host | 8.5 | – | – |
| MiMo-V2.6-FlashXiaomi · open weights · developer price | $0.18 | $0.14 | $0.28 | $0.14GMICloud, BF16 · 7 hosts | 37.9 | – | 46 |
| Seed-2.0-MiniByteDance · typical host price | $0.18 | $0.10 | $0.40 | one host | – | – | – |
| DeepSeek-V4-Flash (0731)DeepSeek · open weights · typical host price | $0.20 | $0.13 | $0.40 | $0.088Relace, FP4 · 27 hosts | 34.3 | – | – |
| Claude Haiku 5.5Anthropic · typical host price | $0.20 | $0.10 | $0.50 | $0.20Amazon Bedrock · 10 hosts | 43.4 | 38 | 240 |
| GPT-6 LunaOpenAI · developer price | $0.20 | $0.10 | $0.50 | $0.20Amazon Bedrock · 6 hosts | 38.1 | 38 | 127 |
| Ministral 3 14BMistral AI · open weights · developer price | $0.20 | $0.20 | $0.20 | $0.20Mistral · 2 hosts | 6.0 | – | 79 |
| Muse EmberMeta · developer priceRequests with more than 200,000 input tokens are billed in full at $0.30 input and $1.50 output. | $0.20 | $0.10 | $0.50 | one host | – | 61 | – |
| gpt-oss-120bOpenAI · open weights · typical host price | $0.20 | $0.10 | $0.50 | $0.060Near AI, FP4 · 25 hosts | 11.6 | – | 259 |
| Qwen3-32BAlibaba · open weights · typical host price | $0.21 | $0.14 | $0.40 | $0.13DeepInfra, FP8 · 3 hosts | 7.3 | – | – |
| Gemma 4 31BGoogle · open weights · typical host price | $0.21 | $0.15 | $0.40 | $0.16CoreWeave, FP4 · 17 hosts | 14.7 | – | 48 |
| Qwen3-30B-A3BAlibaba · open weights · typical host price | $0.21 | $0.12 | $0.50 | one host | 7.6 | – | – |
| Qwen3.8-FlashAlibaba · open weights · typical host price | $0.23 | $0.15 | $0.47 | $0.23Alibaba · 2 hosts | – | – | – |
| GLM-5.3-FlashZ.ai · open weights · developer price | $0.24 | $0.15 | $0.50 | $0.050Relace, FP4 · 40 hosts | 41.8 | – | 79 |
| Nemotron 3.5 Super VLNVIDIA · open weights · typical host price | $0.24 | $0.15 | $0.50 | $0.13Darkbloom, NVFP4 · 10 hosts | 26.5 | – | 411 |
| Hy3Tencent · open weights · typical host price | $0.25 | $0.14 | $0.58 | $0.23Tencent, FP8 · 5 hosts | 25.3 | – | 83 |
| Command R (08-2024)Cohere · typical host price | $0.26 | $0.15 | $0.60 | one host | – | – | – |
| GPT-4o mini (2024-07-18)OpenAI · typical host price | $0.26 | $0.15 | $0.60 | one host | 6.7 | – | – |
| MiniMax M3.1 FlashMiniMax · open weights · developer priceFlat rate, from MiniMax's launch post; on 19 Oct OpenRouter only offered a free route | $0.26 | $0.15 | $0.60 | $0.20GMICloud, FP8 · 13 hosts | 42.4 | 40 | 159 |
| Mistral Small 4Mistral AI · open weights · developer price | $0.26 | $0.15 | $0.60 | $0.26Mistral · 3 hosts | 11.3 | – | 176 |
| Solar Pro 4Upstage · typical host price | $0.26 | $0.15 | $0.60 | $0.26Upstage · 2 hosts | 26.6 | – | 119 |
| Hy3 PreviewTencent · open weights · typical host price | $0.29 | $0.18 | $0.60 | one host | 22.7 | – | – |
| Qwen3-Coder-NextAlibaba · open weights · typical host price | $0.29 | $0.12 | $0.80 | one host | 9.2 | – | 77 |
| Qwen3-235B-A22B-Instruct-2507Alibaba · open weights · typical host price | $0.30 | $0.15 | $0.75 | $0.21DeepInfra, FP8 · 3 hosts | 12.0 | – | – |
| DeepSeek-V3.2-ExpDeepSeek · open weights · typical host price | $0.30 | $0.27 | $0.41 | one host | 16.6 | – | – |
| Qwen3.6-35B-A3BAlibaba · open weights · typical host price | $0.31 | $0.10 | $0.95 | $0.037PrimeIntellect, FP8 · 11 hosts | 18.2 | – | 135 |
| DeepSeek-V3.2DeepSeek · open weights · typical host price | $0.32 | $0.28 | $0.42 | $0.23GMICloud, FP8 · 7 hosts | 21.5 | – | – |
| GLM-4.5-AirZ.ai · open weights · typical host price | $0.32 | $0.14 | $0.86 | $0.31Novita, BF16 · 3 hosts | 11.1 | – | – |
| DeepSeek-V4-Flash-Vision-ExpDeepSeek · open weights · typical host price | $0.32 | $0.22 | $0.65 | $0.32DeepInfra, FP8 · 2 hosts | 34.8 | – | 191 |
| Qwen3-VL-30B-A3B-InstructAlibaba · open weights · typical host price | $0.33 | $0.20 | $0.70 | $0.26DeepInfra, FP8 · 3 hosts | 7.9 | – | – |
| Qwen3-Next-80B-A3B-InstructAlibaba · open weights · typical host price | $0.34 | $0.090 | $1.10 | $0.34DeepInfra, FP8 · 2 hosts | 9.6 | – | 121 |
| Qwen3.5-35B-A3BAlibaba · open weights · typical host price | $0.36 | $0.15 | $1.00 | $0.24Darkbloom, FP4 · 7 hosts | 19.3 | – | – |
| Qwen2.5-72B-InstructAlibaba · open weights · typical host price | $0.37 | $0.36 | $0.40 | one host | 7.7 | – | – |
| Qwen3-VL-8B-InstructAlibaba · open weights · typical host price | $0.38 | $0.25 | $0.75 | one host | 7.3 | – | – |
| Qwen-PlusAlibaba · typical host price | $0.39 | $0.26 | $0.78 | one host | – | – | – |
| Mistral Small 3.1 24BMistral AI · open weights · typical host price | $0.40 | $0.35 | $0.56 | one host | 7.1 | – | – |
| Qwen3.6-FlashAlibaba · typical host price | $0.42 | $0.19 | $1.12 | one host | – | – | – |
| DeepSeek-V3-0324DeepSeek · open weights · typical host price | $0.44 | $0.25 | $1.00 | $0.44SiliconFlow, FP8 · 2 hosts | 9.7 | – | – |
| GLM-4.6VZ.ai · open weights · typical host price | $0.45 | $0.30 | $0.90 | $0.45Novita, BF16 · 2 hosts | 11.2 | – | – |
| GPT-5.6 LunaOpenAI · developer price | $0.45 | $0.20 | $1.20 | $0.45Amazon Bedrock · 6 hosts | 37.3 | – | – |
| DeepSeek-V3DeepSeek · open weights · typical host price | $0.45 | $0.26 | $1.03 | $0.45StreamLake · 2 hosts | 8.5 | – | – |
| DeepSeek-V3.1DeepSeek · open weights · typical host price | $0.45 | $0.27 | $1.00 | $0.42DeepInfra, FP4 · 4 hosts | 13.7 | – | – |
| DeepSeek-V3.1-TerminusDeepSeek · open weights · typical host price | $0.45 | $0.27 | $1.00 | one host | 14.8 | – | – |
| GPT-5.4 nanoOpenAI · developer price | $0.46 | $0.20 | $1.25 | $0.46Azure · 3 hosts | 20.7 | – | – |
| Qwen3-VL-30B-A3B-ThinkingAlibaba · open weights · typical host price | $0.47 | $0.29 | $1.00 | one host | 9.5 | – | – |
| Qwen3-Coder-480B-A35BAlibaba · open weights · typical host price | $0.47 | $0.30 | $1.00 | $0.47DeepInfra, FP4 · 2 hosts | 11.9 | – | – |
| Llama 4 MaverickMeta · open weights · typical host price | $0.51 | $0.35 | $1.00 | $0.51Parasail, FP8 · 2 hosts | 9.3 | – | 46 |
| MiniMax-M2.5MiniMax · open weights · typical host price | $0.52 | $0.29 | $1.20 | $0.44Venice · 5 hosts | 22.8 | – | – |
| DeepSeek-V4.1-FlashDeepSeek · open weights · developer priceHalf price off-peak ($0.15 / $0.60) | $0.52 | $0.30 | $1.20 | $0.12Morph, FP8 · 34 hosts | 39.5 | 49 | 369 |
| DeepSeek-V4.1-Flash-1017DeepSeek · open weights · developer priceHalf price off-peak ($0.15 / $0.60) | $0.52 | $0.30 | $1.20 | $0.11Sail Research, FP4 · 45 hosts | 45.1 | 37 | 212 |
| MiniMax-M2.7MiniMax · open weights · developer price | $0.52 | $0.30 | $1.20 | $0.37GMICloud, FP8 · 2 hosts | 22.8 | – | – |
| MiniMax-M3MiniMax · open weights · developer priceShown as a permanent 50% off $0.60 / $2.40 | $0.52 | $0.30 | $1.20 | $0.41CoreWeave, FP4 · 9 hosts | 29.2 | – | 146 |
| Muse Glimmer 30BMeta · open weights · typical host price | $0.52 | $0.30 | $1.20 | one host | – | – | – |
| MiMo-V2.5-ProXiaomi · open weights · typical host price | $0.54 | $0.43 | $0.87 | $0.54AtlasCloud, FP8 · 2 hosts | 26.0 | – | – |
| MiMo-V2.6-ProXiaomi · open weights · developer price | $0.54 | $0.43 | $0.87 | $0.41GMICloud, BF16 · 5 hosts | 46.3 | – | 43 |
| Qwen3.7-PlusAlibaba · typical host price | $0.56 | $0.32 | $1.28 | $0.56Alibaba · 2 hosts | 25.2 | – | 56 |
| Gemini 3.1 Flash-LiteGoogle · typical host price | $0.56 | $0.25 | $1.50 | $0.56Google · 4 hosts | – | – | – |
| Gemini 3.1 Flash-Lite PreviewGoogle · typical host price | $0.56 | $0.25 | $1.50 | one host | 15.6 | – | – |
| Qwen3.5-Plus (2026-02-15)Alibaba · typical host price | $0.58 | $0.26 | $1.56 | one host | – | – | – |
| GLM-5.3 FlashX LegalZ.ai · typical host price | $0.59 | $0.37 | $1.25 | one host | – | 29 | – |
| Beam 1.0 501B A23BReflection · open weights · typical host price | $0.60 | $0.40 | $1.20 | $0.44Nebius, FP4 · 5 hosts | 30.3 | – | 323 |
| Qwen3-VL-235B-A22B-InstructAlibaba · open weights · typical host price | $0.63 | $0.21 | $1.90 | $0.37DeepInfra, FP8 · 3 hosts | 9.9 | – | – |
| Inkling SmallThinking Machines · open weights · typical host price | $0.64 | $0.45 | $1.20 | one host | 25.7 | – | 181 |
| Gemma 2 27BGoogle · open weights · typical host price | $0.65 | $0.65 | $0.65 | one host | – | – | – |
| GPT-5 miniOpenAI · typical host price | $0.69 | $0.25 | $2.00 | $0.69Azure · 3 hosts | 20.6 | – | – |
| GPT-5.1-Codex-MiniOpenAI · typical host price | $0.69 | $0.25 | $2.00 | one host | 20.4 | – | – |
| GPT-4.1 miniOpenAI · typical host price | $0.70 | $0.40 | $1.60 | $0.70Azure · 3 hosts | 10.2 | – | – |
| Llama 3.3 70B InstructMeta · open weights · typical host price | $0.71 | $0.71 | $0.71 | $0.16DeepInfra, FP8 · 9 hosts | 7.7 | – | 98 |
| Qwen3.5-122B-A10BAlibaba · open weights · typical host price | $0.72 | $0.26 | $2.08 | $0.72Alibaba · 4 hosts | 17.7 | – | 144 |
| Llama 3.1 70B InstructMeta · open weights · typical host price | $0.72 | $0.72 | $0.72 | one host | 6.6 | – | – |
| Qwen3.6-PlusAlibaba · typical host price | $0.73 | $0.33 | $1.95 | one host | 27.0 | – | – |
| Qwen3.5-27BAlibaba · open weights · typical host price | $0.74 | $0.27 | $2.16 | $0.54Alibaba · 7 hosts | 22.9 | – | – |
| Qwen2.5-Coder-32B-InstructAlibaba · open weights · typical host price | $0.74 | $0.66 | $1.00 | one host | 6.7 | – | – |
| Mistral Large 3Mistral AI · developer price | $0.75 | $0.50 | $1.50 | $0.75Mistral · 2 hosts | 9.3 | – | – |
| Devstral 2Mistral AI · open weights · typical host price | $0.80 | $0.40 | $2.00 | one host | 8.6 | – | – |
| MiMo-V2.5Xiaomi · open weights · typical host price | $0.80 | $0.40 | $2.00 | one host | 21.8 | – | – |
| Mistral Medium 3Mistral AI · typical host price | $0.80 | $0.40 | $2.00 | $0.80Mistral · 2 hosts | 9.0 | – | – |
| Mistral Medium 3.1Mistral AI · typical host price | $0.80 | $0.40 | $2.00 | $0.80Mistral · 2 hosts | 9.2 | – | – |
| Gemini 2.5 FlashGoogle · developer price | $0.85 | $0.30 | $2.50 | one host | 13.1 | – | – |
| Gemini 3.5 Flash-LiteGoogle · developer price | $0.85 | $0.30 | $2.50 | $0.85Google · 4 hosts | 22.2 | 24 | 398 |
| Nova 2 LiteAmazon · typical host price | $0.85 | $0.30 | $2.50 | $0.85Amazon Bedrock · 2 hosts | 13.4 | – | 233 |
| Qwen2.5-VL-72B-InstructAlibaba · open weights · typical host price | $0.85 | $0.80 | $1.00 | one host | – | – | – |
| GLM-4.6Z.ai · open weights · typical host price | $0.88 | $0.50 | $2.00 | $0.76Venice, FP4 · 4 hosts | 18.5 | – | – |
| DeepSeek-R1-0528DeepSeek · open weights · typical host price | $0.92 | $0.50 | $2.18 | $0.91DeepInfra, FP4 · 3 hosts | 13.1 | – | – |
| GLM-4.7Z.ai · open weights · typical host price | $1.00 | $0.60 | $2.20 | $0.74DeepInfra, FP4 · 5 hosts | 22.2 | – | – |
| Kimi K2 (0711)Moonshot AI · open weights · typical host price | $1.00 | $0.57 | $2.30 | one host | 12.7 | – | – |
| Qwen3.6-27BAlibaba · open weights · typical host price | $1.02 | $0.30 | $3.20 | $0.69Chutes, FP8 · 7 hosts | 21.4 | – | – |
| Kimi K2 (0905)Moonshot AI · open weights · typical host price | $1.07 | $0.60 | $2.50 | one host | 15.3 | – | – |
| Kimi K2 ThinkingMoonshot AI · open weights · typical host price | $1.07 | $0.60 | $2.50 | one host | 22.0 | – | – |
| Gemini 3 Flash PreviewGoogle · typical host price | $1.12 | $0.50 | $3.00 | $1.12Google · 2 hosts | 26.3 | – | – |
| Qwen3.8-27BAlibaba · open weights · developer price | $1.12 | $0.50 | $3.00 | $0.13PrimeIntellect, FP8 · 25 hosts | 33.7 | – | 46 |
| Qwen3-235B-A22B-Thinking-2507Alibaba · open weights · typical host price | $1.21 | $0.45 | $3.50 | one host | 12.7 | – | – |
| Grok Build 0.1xAI · developer price | $1.25 | $1.00 | $2.00 | $1.25xAI · 2 hosts | 27.2 | – | – |
| Hy4 PreviewTencent · open weights · typical host price | $1.25 | $0.83 | $2.50 | $1.25AtlasCloud, FP8 · 6 hosts | – | – | – |
| Qwen3.5-397B-A17BAlibaba · open weights · typical host price | $1.29 | $0.55 | $3.50 | $0.88Alibaba · 9 hosts | 21.4 | – | 56 |
| Nova Pro 1.0Amazon · typical host price | $1.40 | $0.80 | $3.20 | $1.40Amazon Bedrock · 2 hosts | 7.0 | – | – |
| Step 5 PreviewStepFun · open weights · typical host price | $1.43 | $1.00 | $2.70 | $1.43Featherless, FP4 · 2 hosts | – | – | – |
| GLM-5Z.ai · open weights · typical host price | $1.55 | $1.00 | $3.20 | $0.93GMICloud, FP8 · 9 hosts | 27.9 | – | – |
| Grok 4.20xAI · typical host price | $1.56 | $1.25 | $2.50 | $1.56xAI · 2 hosts | 25.7 | – | – |
| Grok 4.3xAI · developer price | $1.56 | $1.25 | $2.50 | $1.56xAI · 2 hosts | 24.9 | – | – |
| GPT-5.4 miniOpenAI · developer price | $1.69 | $0.75 | $4.50 | $1.69Azure · 3 hosts | 24.1 | – | – |
| Kimi K2.6Moonshot AI · open weights · developer price | $1.71 | $0.95 | $4.00 | $0.74Chutes, INT4 · 16 hosts | 27.0 | – | – |
| Kimi K2.7 CodeMoonshot AI · open weights · developer price | $1.71 | $0.95 | $4.00 | $1.20CoreWeave, INT4 · 11 hosts | 25.8 | – | 54 |
| InklingThinking Machines · open weights · typical host price | $1.76 | $1.00 | $4.05 | one host | 25.0 | – | 232 |
| DeepSeek-V4-Pro (0423)DeepSeek · open weights · typical host price | $1.77 | $1.42 | $2.83 | $1.05Relace, FP4 · 13 hosts | 30.4 | 27 | – |
| GLM-5-TurboZ.ai · typical host price | $1.90 | $1.20 | $4.00 | one host | 26.6 | – | – |
| GLM-5V-TurboZ.ai · typical host price | $1.90 | $1.20 | $4.00 | one host | 23.5 | 12 | – |
| DeepSeek-V4-Pro (0813)DeepSeek · open weights · developer priceHalf price off-peak ($0.66 / $1.98) | $1.98 | $1.32 | $3.96 | $0.87Alibaba · 24 hosts | 36.0 | – | 96 |
| Claude Haiku 4.5Anthropic · developer price | $2.00 | $1.00 | $5.00 | $2.00Amazon Bedrock · 8 hosts | 16.9 | 7 | – |
| GPT-6.1 TerraOpenAI · developer priceShort-context rate, up to 272K input tokens | $2.00 | $1.00 | $5.00 | $2.00Azure · 5 hosts | – | 77 | – |
| Muse Spark 1.1Meta · typical host price | $2.00 | $1.25 | $4.25 | one host | 33.7 | – | – |
| Muse Spark 1.2Meta · typical host price | $2.00 | $1.25 | $4.25 | one host | 39.6 | – | – |
| Muse Spark 1.3Meta · developer priceAlso $0.10 / $0.20 if Meta may train on your data | $2.00 | $1.25 | $4.25 | one host | 48.1 | 58 | 254 |
| Muse Spark OpenMeta · open weights · typical host price | $2.00 | $1.25 | $4.25 | $2.00BaseTen, FP4 · 5 hosts | 38.9 | – | 142 |
| GLM-5.1Z.ai · open weights · typical host price | $2.04 | $1.33 | $4.18 | $1.40Chutes, FP8 · 12 hosts | 26.1 | – | – |
| Mistral Large 4Mistral AI · open weights · developer price | $2.06 | $1.36 | $4.18 | $1.05Mistral · 2 hosts | 38.4 | 24 | 60 |
| GLM-5.2Z.ai · open weights · developer price | $2.15 | $1.40 | $4.40 | $0.87DeepInfra, FP4 · 28 hosts | 33.7 | – | – |
| GLM-5.3Z.ai · open weights · developer price | $2.15 | $1.40 | $4.40 | $0.40Wafer, FP4 · 43 hosts | 44.8 | 35 | 87 |
| Qwen3.7-MaxAlibaba · typical host price | $2.21 | $1.48 | $4.42 | $2.21Alibaba · 2 hosts | 29.5 | – | – |
| Qwen4-MaxAlibaba · OpenRouter price | $2.25 | $1.50 | $4.50 | one host | – | 53 | – |
| Gemini 3.6 FlashGoogle · developer priceIntroductory price: $0.75 in, $3.75 out until 31 Dec 2026 | $3.00 | $1.50 | $7.50 | $1.50Google · 3 hosts | 34.0 | – | – |
| Gemini 3.7 FlashGoogle · developer priceIntroductory price: $0.75 in, $3.75 out until 31 Dec 2026 | $3.00 | $1.50 | $7.50 | $1.50Google · 3 hosts | 39.6 | – | – |
| Gemini 3.8 FlashGoogle · developer priceIntroductory price: $0.75 in, $3.75 out until 31 Dec 2026 | $3.00 | $1.50 | $7.50 | $1.50Google · 3 hosts | 40.9 | 57 | 141 |
| Gemini 4 HydrogenGoogle · developer priceSame price across the full 1M context. | $3.00 | $1.50 | $7.50 | one host | – | 60 | – |
| Grok 4.5xAI · developer price | $3.00 | $2.00 | $6.00 | $3.00xAI · 2 hosts | 38.8 | – | – |
| Grok 4.6xAI · developer price | $3.00 | $2.00 | $6.00 | $3.00xAI · 4 hosts | 44.3 | – | – |
| Grok 4.7xAI · developer price | $3.00 | $2.00 | $6.00 | $3.00Amazon Bedrock · 4 hosts | 46.4 | 39 | 74 |
| Mistral Large 2 (2407)Mistral AI · typical host price | $3.00 | $2.00 | $6.00 | $3.00Mistral · 3 hosts | 6.8 | – | – |
| Mistral Medium 3.5Mistral AI · developer price | $3.00 | $1.50 | $7.50 | $3.00Mistral · 2 hosts | 14.2 | 15 | 157 |
| Mixtral 8x22B InstructMistral AI · open weights · typical host price | $3.00 | $2.00 | $6.00 | $3.00Mistral · 2 hosts | 5.7 | – | – |
| Qwen3.8-2.4T-A95BAlibaba · open weights · developer price | $3.00 | $2.00 | $6.00 | $3.00Alibaba · 4 hosts | 39.9 | – | 36 |
| Qwen3.8-Max (0902)Alibaba · developer price | $3.00 | $2.00 | $6.00 | $3.00Alibaba · 2 hosts | 45.4 | 44 | 37 |
| Gemini 3.5 FlashGoogle · developer price | $3.38 | $1.50 | $9.00 | $3.38Google · 3 hosts | 33.6 | – | – |
| GPT-5OpenAI · typical host price | $3.44 | $1.25 | $10.00 | $3.44Azure · 3 hosts | 23.0 | – | – |
| GPT-5.1OpenAI · typical host price | $3.44 | $1.25 | $10.00 | $3.44Azure · 3 hosts | 24.7 | – | – |
| GPT-5.1-CodexOpenAI · typical host price | $3.44 | $1.25 | $10.00 | one host | 23.7 | – | – |
| Gemini 2.5 ProGoogle · developer price | $3.44 | $1.25 | $10.00 | one host | 16.1 | – | – |
| Gemini 2.5 Pro Preview (06-05)Google · typical host price | $3.44 | $1.25 | $10.00 | one host | – | – | – |
| GPT-4.1OpenAI · typical host price | $3.50 | $2.00 | $8.00 | $3.50Azure · 3 hosts | 12.7 | – | – |
| o3OpenAI · typical host price | $3.50 | $2.00 | $8.00 | one host | 20.2 | – | 155 |
| Kimi K3.1 ProMoonshot AI · open weights · typical host price | $3.81 | $1.75 | $10.00 | $2.38Relace, FP4 · 22 hosts | – | 67 | – |
| Claude Sonnet 5Anthropic · developer price | $4.00 | $2.00 | $10.00 | $4.00Amazon Bedrock · 10 hosts | 38.2 | – | – |
| Claude Sonnet 5.5Anthropic · developer price | $4.00 | $2.00 | $10.00 | $4.00Amazon Bedrock · 10 hosts | 56.0 | 60 | 137 |
| GPT-6 SolOpenAI · developer price | $4.00 | $2.00 | $10.00 | $4.00Amazon Bedrock · 6 hosts | 47.6 | 66 | – |
| GPT-6.1 SolOpenAI · developer price | $4.00 | $2.00 | $10.00 | $4.00Amazon Bedrock · 7 hosts | 51.8 | 78 | 53 |
| GPT-6.2 SolOpenAI · OpenRouter price | $4.00 | $2.00 | $10.00 | one host | – | 84 | – |
| Gemini 4 ArgonGoogle · typical host price | $4.00 | $2.00 | $10.00 | $4.00Google · 3 hosts | 52.6 | 70 | – |
| Command ACohere · open weights · typical host price | $4.38 | $2.50 | $10.00 | one host | 7.0 | – | 59 |
| Command R+ (08-2024)Cohere · typical host price | $4.38 | $2.50 | $10.00 | one host | – | – | – |
| GPT-4o (2024-08-06)OpenAI · typical host price | $4.38 | $2.50 | $10.00 | $4.38Azure · 2 hosts | 7.7 | – | – |
| GPT-5.6 TerraOpenAI · developer price | $4.50 | $2.00 | $12.00 | $4.50Azure · 5 hosts | 42.1 | – | 140 |
| Gemini 3.1 Pro PreviewGoogle · developer price | $4.50 | $2.00 | $12.00 | $4.50Google · 2 hosts | 29.7 | 44 | 130 |
| Gemini 3.1 Pro Preview (custom tools)Google · typical host price | $4.50 | $2.00 | $12.00 | one host | – | – | – |
| GPT-5.2OpenAI · typical host price | $4.81 | $1.75 | $14.00 | $4.81Azure · 2 hosts | 30.4 | – | – |
| GPT-5.2-CodexOpenAI · typical host price | $4.81 | $1.75 | $14.00 | one host | 28.5 | – | – |
| GPT-5.3-CodexOpenAI · typical host price | $4.81 | $1.75 | $14.00 | $4.81Azure · 2 hosts | 32.5 | – | 119 |
| GPT-5.4OpenAI · developer price | $5.62 | $2.50 | $15.00 | $5.62Azure · 5 hosts | 39.0 | – | – |
| Claude Sonnet 4.5Anthropic · typical host price | $6.00 | $3.00 | $15.00 | $6.00Amazon Bedrock · 7 hosts | 20.7 | – | – |
| Claude Sonnet 4.6Anthropic · typical host price | $6.00 | $3.00 | $15.00 | $6.00Amazon Bedrock · 9 hosts | 30.1 | – | – |
| Gemini 4 NeonGoogle · OpenRouter price | $6.00 | $3.00 | $15.00 | one host | – | 60 | – |
| Kimi K3Moonshot AI · open weights · developer price | $6.00 | $3.00 | $15.00 | $2.47Relace, FP4 · 24 hosts | 43.6 | 46 | 41 |
| GPT-4o (2024-05-13)OpenAI · typical host price | $7.50 | $5.00 | $15.00 | one host | 7.3 | – | – |
| Claude Opus 5.5Anthropic · developer price | $8.00 | $4.00 | $20.00 | $8.00Amazon Bedrock · 10 hosts | 57.6 | 80 | 96 |
| GPT-5.6 SolOpenAI · developer pricePromotional price until at least 21 Nov 2026 | $8.00 | $4.00 | $20.00 | $8.00Amazon Bedrock · 6 hosts | 47.0 | – | – |
| GPT-5.6 Sol ProOpenAI · typical host price | $8.00 | $4.00 | $20.00 | $8.00Azure · 3 hosts | – | – | – |
| Gemini 4 CarbonGoogle · developer priceIntroductory price: $2.00 in, $10.00 out until 30 Nov 2026; Standard prices apply from 1 December 2026. Output pricing includes thinking tokens. | $8.00 | $4.00 | $20.00 | one host | – | 50 | – |
| Muse WonderMeta · developer priceStandard Wonder. Muse Wonder Full is offered separately through Muse Code subscriptions. | $8.00 | $4.00 | $20.00 | one host | 52.4 | 66 | 153 |
| Claude Opus 4.5Anthropic · typical host price | $10.00 | $5.00 | $25.00 | $10.00Amazon Bedrock · 6 hosts | 29.1 | – | – |
| Claude Opus 4.6Anthropic · typical host price | $10.00 | $5.00 | $25.00 | $10.00Amazon Bedrock · 6 hosts | 31.9 | – | – |
| Claude Opus 4.7Anthropic · typical host price | $10.00 | $5.00 | $25.00 | $10.00Amazon Bedrock · 8 hosts | 40.7 | – | – |
| Claude Opus 4.8Anthropic · typical host price | $10.00 | $5.00 | $25.00 | $10.00Amazon Bedrock · 10 hosts | 41.8 | – | – |
| Claude Opus 5Anthropic · developer price | $10.00 | $5.00 | $25.00 | $10.00Amazon Bedrock · 10 hosts | 50.8 | – | – |
| GPT-5.5OpenAI · developer price | $11.25 | $5.00 | $30.00 | $11.25Azure · 5 hosts | 38.4 | – | – |
| Claude Fable 5.2Anthropic · developer price | $16.00 | $8.00 | $40.00 | $16.00Amazon Bedrock · 4 hosts | – | 88 | – |
| GPT-6.2 AstraOpenAI · OpenRouter price | $16.00 | $8.00 | $40.00 | one host | – | 86 | – |
| Claude Fable 5Anthropic · developer price | $20.00 | $10.00 | $50.00 | $20.00Amazon Bedrock · 6 hosts | 49.6 | – | – |
| Claude Fable 5.1Anthropic · developer price | $20.00 | $10.00 | $50.00 | $20.00Amazon Bedrock · 4 hosts | 53.4 | 64 | 72 |
| GPT-6 AstraOpenAI · developer price | $20.00 | $10.00 | $50.00 | $20.00Amazon Bedrock · 6 hosts | 52.7 | 76 | 72 |
| Claude Opus 4.1Anthropic · typical host price | $30.00 | $15.00 | $75.00 | one host | 22.8 | – | – |
| o3-proOpenAI · typical host price | $35.00 | $20.00 | $80.00 | one host | 21.9 | – | – |
| GPT-5.4 ProOpenAI · typical host price | $67.50 | $30.00 | $180.00 | $67.50Azure · 2 hosts | – | – | – |
The cheapest model at each level
Grouped by the Artificial Analysis Intelligence Index. The cheapest model in a group is rarely the most capable one, but it is often close.
| Intelligence | Cheapest | Price | Most capable | Price |
|---|---|---|---|---|
| Frontier: 50 or more | Claude Sonnet 5.5Anthropic · 56.0 | $4.00 | Claude Opus 5.5Anthropic · 57.6 | $8.00 |
| Strong: 40–50 | Claude Haiku 5.5Anthropic · 43.4 | $0.20 | Claude Fable 5Anthropic · 49.6 | $20.00 |
| Capable: 30–40 | MiMo-V2.6-FlashXiaomi · 37.9 | $0.18 | Qwen3.8-2.4T-A95BAlibaba · 39.9 | $3.00 |
| Everyday: 20–30 | DeepSeek-V4-Flash (0423)DeepSeek · 24.4 | $0.15 | Gemini 3.1 Pro PreviewGoogle · 29.7 | $4.50 |
| Light: under 20 | Gemma 3 4BGoogle · 4.8 | $0.042 | Qwen3.5-35B-A3BAlibaba · 19.3 | $0.36 |
Price against intelligence
What each model costs for what it can do
Further up is more capable; further left is cheaper. Green models are the best value: none is beaten on both.
Best valueOther models (hover for names)
Each dot is a model on sale today. Green dots are the best-value models: for each, no other model is both cheaper and more capable. Those scoring 20 or more are named; the cheap end is too crowded for labels, so every best-value model is also listed under Best value. Price is on a log scale.
Best and worst value
Best value
Each one is the cheapest way to reach its level of intelligence.
- Claude Opus 5.5 57.6 for $8.00
- Claude Sonnet 5.5 56.0 for $4.00
- Muse Spark 1.3 48.1 for $2.00
- MiMo-V2.6-Pro 46.3 for $0.54
- DeepSeek-V4.1-Flash-1017 45.1 for $0.52
- Claude Haiku 5.5 43.4 for $0.20
- MiMo-V2.6-Flash 37.9 for $0.18
- DeepSeek-V4-Flash (0423) 24.4 for $0.15
- Gemma 4 26B A4B 16.7 for $0.14
- Qwen3.5-9B 13.3 for $0.11
- Granite 4.2 8B 11.1 for $0.082
- gpt-oss-20b 10.0 for $0.058
- Gemma 3 4B 4.8 for $0.042
Expensive for what you get
At least 3× the price of the cheapest model that is as capable, on the Intelligence Index alone. A model can still be worth it for a task the index doesn't measure: check its model page.
- o3-pro $35.00: 233× DeepSeek-V4-Flash (0423) ($0.15, 24.4)
- Claude Opus 4.1 $30.00: 200× DeepSeek-V4-Flash (0423) ($0.15, 24.4)
- GPT-4o (2024-05-13) $7.50: 130× gpt-oss-20b ($0.058, 10.0)
- Command A $4.38: 76× gpt-oss-20b ($0.058, 10.0)
- GPT-4o (2024-08-06) $4.38: 76× gpt-oss-20b ($0.058, 10.0)
- Claude Opus 4.5 $10.00: 57× MiMo-V2.6-Flash ($0.18, 37.9)
- Claude Opus 4.6 $10.00: 57× MiMo-V2.6-Flash ($0.18, 37.9)
- GPT-5.5 $11.25: 56× Claude Haiku 5.5 ($0.20, 43.4)
- Mistral Large 2 (2407) $3.00: 52× gpt-oss-20b ($0.058, 10.0)
- Mixtral 8x22B Instruct $3.00: 52× gpt-oss-20b ($0.058, 10.0)
- Claude Opus 4.7 $10.00: 50× Claude Haiku 5.5 ($0.20, 43.4)
- Claude Opus 4.8 $10.00: 50× Claude Haiku 5.5 ($0.20, 43.4)
Price per token isn't cost per task
A model that writes three times as many tokens costs three times as much, whatever its price list says. These are the costs we measured running our own benchmarks, at each model's cheapest setting. Each table has the models that benchmark ran.
CatalogBench: turning a sparse product feed into listings (36 models)
| Model | Blended price | Rank by price | Measured cost per 10,000 listings | Rank by cost |
|---|---|---|---|---|
| MiniMax M3.1 Flash | $0.26 | 5 | $0.00 | 1 ▲ cheaper than its price suggests |
| GPT-6 Luna | $0.20 | 3 | $4.77 | 2 |
| Particle-1 | $0.16 | 1 | $11.72 | 3 |
| Claude Haiku 5.5 | $0.20 | 2 | $12.08 | 4 |
| DeepSeek-V4.1-Flash-1017 | $0.52 | 7 | $15.42 | 5 |
| Muse Ember | $0.20 | 4 | $20.40 | 6 |
| Gemini 3.5 Flash-Lite | $0.85 | 9 | $20.47 | 7 |
| Mistral Large 4 | $2.06 | 14 | $32.77 | 8 ▲ cheaper than its price suggests |
| GPT-6.1 Terra | $2.00 | 12 | $35.60 | 9 ▲ cheaper than its price suggests |
| GLM-5.3 FlashX Legal | $0.59 | 8 | $50.35 | 10 |
| DeepSeek-V4.1-Flash | $0.52 | 6 | $60.70 | 11 ▼ dearer than its price suggests |
| Claude Haiku 4.5 | $2.00 | 11 | $61.38 | 12 |
| Qwen4-Max | $2.25 | 15 | $68.53 | 13 |
| GLM-5V-Turbo | $1.90 | 10 | $70.11 | 14 ▼ dearer than its price suggests |
| Mistral Medium 3.5 | $3.00 | 19 | $80.05 | 15 ▲ cheaper than its price suggests |
| GPT-6.2 Sol | $4.00 | 25 | $82.68 | 16 ▲ cheaper than its price suggests |
| GPT-6.1 Sol | $4.00 | 24 | $83.84 | 17 ▲ cheaper than its price suggests |
| GPT-6 Sol | $4.00 | 23 | $85.18 | 18 ▲ cheaper than its price suggests |
| Gemini 3.8 Flash | $3.00 | 16 | $102.92 | 19 ▼ dearer than its price suggests |
| Gemini 4 Argon | $4.00 | 26 | $109.21 | 20 ▲ cheaper than its price suggests |
| Muse Spark 1.3 | $2.00 | 13 | $164.18 | 21 ▼ dearer than its price suggests |
| Claude Sonnet 5.5 | $4.00 | 22 | $171.91 | 22 |
| Gemini 4 Carbon | $8.00 | 31 | $213.47 | 23 ▲ cheaper than its price suggests |
| Gemini 4 Hydrogen | $3.00 | 17 | $219.33 | 24 ▼ dearer than its price suggests |
| Grok 4.7 | $3.00 | 18 | $237.19 | 25 ▼ dearer than its price suggests |
| Qwen3.8-Max (0902) | $3.00 | 20 | $255.72 | 26 ▼ dearer than its price suggests |
| GPT-6.2 Astra | $16.00 | 34 | $345.26 | 27 ▲ cheaper than its price suggests |
| Gemini 3.1 Pro Preview | $4.50 | 27 | $354.64 | 28 |
| Muse Wonder | $8.00 | 32 | $360.73 | 29 ▲ cheaper than its price suggests |
| GPT-6 Astra | $20.00 | 36 | $371.20 | 30 ▲ cheaper than its price suggests |
| Claude Opus 5.5 | $8.00 | 30 | $400.06 | 31 |
| Kimi K3.1 Pro | $3.81 | 21 | $454.38 | 32 ▼ dearer than its price suggests |
| Kimi K3 | $6.00 | 29 | $468.95 | 33 ▼ dearer than its price suggests |
| Gemini 4 Neon | $6.00 | 28 | $652.78 | 34 ▼ dearer than its price suggests |
| Claude Fable 5.2 | $16.00 | 33 | $875.39 | 35 |
| Claude Fable 5.1 | $20.00 | 35 | $991.28 | 36 |
DeckBench: turning an analysis into a deck (35 models)
| Model | Blended price | Rank by price | Measured cost per 100 decks | Rank by cost |
|---|---|---|---|---|
| MiniMax M3.1 Flash | $0.26 | 5 | $0.00 | 1 ▲ cheaper than its price suggests |
| Particle-1 | $0.16 | 1 | $0.86 | 2 |
| GPT-6 Luna | $0.20 | 3 | $0.86 | 3 |
| Claude Haiku 5.5 | $0.20 | 2 | $0.91 | 4 |
| Muse Ember | $0.20 | 4 | $1.12 | 5 |
| Gemini 3.5 Flash-Lite | $0.85 | 8 | $1.26 | 6 |
| Mistral Large 4 | $2.06 | 13 | $1.48 | 7 ▲ cheaper than its price suggests |
| DeepSeek-V4.1-Flash-1017 | $0.52 | 6 | $2.55 | 8 |
| GLM-5.3 FlashX Legal | $0.59 | 7 | $3.59 | 9 |
| Mistral Medium 3.5 | $3.00 | 18 | $4.48 | 10 ▲ cheaper than its price suggests |
| Claude Haiku 4.5 | $2.00 | 10 | $4.61 | 11 |
| Muse Spark 1.3 | $2.00 | 12 | $4.63 | 12 |
| DeepSeek-V4-Pro (0423) | $1.77 | 9 | $4.65 | 13 ▼ dearer than its price suggests |
| Gemini 3.8 Flash | $3.00 | 15 | $6.13 | 14 |
| GPT-6.1 Terra | $2.00 | 11 | $6.72 | 15 ▼ dearer than its price suggests |
| GPT-6.2 Sol | $4.00 | 24 | $7.75 | 16 ▲ cheaper than its price suggests |
| Gemini 4 Argon | $4.00 | 25 | $11.47 | 17 ▲ cheaper than its price suggests |
| Gemini 3.1 Pro Preview | $4.50 | 26 | $13.55 | 18 ▲ cheaper than its price suggests |
| Qwen3.8-Max (0902) | $3.00 | 19 | $14.53 | 19 |
| Claude Sonnet 5.5 | $4.00 | 21 | $15.27 | 20 |
| GPT-6.1 Sol | $4.00 | 23 | $15.29 | 21 |
| Gemini 4 Carbon | $8.00 | 30 | $16.38 | 22 ▲ cheaper than its price suggests |
| Qwen4-Max | $2.25 | 14 | $17.89 | 23 ▼ dearer than its price suggests |
| Muse Wonder | $8.00 | 31 | $18.21 | 24 ▲ cheaper than its price suggests |
| GPT-6.2 Astra | $16.00 | 33 | $19.57 | 25 ▲ cheaper than its price suggests |
| GPT-6 Sol | $4.00 | 22 | $20.97 | 26 ▼ dearer than its price suggests |
| Gemini 4 Hydrogen | $3.00 | 16 | $24.98 | 27 ▼ dearer than its price suggests |
| Grok 4.7 | $3.00 | 17 | $25.24 | 28 ▼ dearer than its price suggests |
| Claude Opus 5.5 | $8.00 | 29 | $26.10 | 29 |
| Kimi K3 | $6.00 | 28 | $45.37 | 30 |
| Kimi K3.1 Pro | $3.81 | 20 | $54.07 | 31 ▼ dearer than its price suggests |
| Gemini 4 Neon | $6.00 | 27 | $77.16 | 32 ▼ dearer than its price suggests |
| Claude Fable 5.1 | $20.00 | 34 | $80.66 | 33 |
| GPT-6 Astra | $20.00 | 35 | $109.21 | 34 |
| Claude Fable 5.2 | $16.00 | 32 | $139.57 | 35 ▼ dearer than its price suggests |
SurveyBench: planning a user survey and analysing its responses (26 models)
| Model | Blended price | Rank by price | Measured cost per 100 surveys | Rank by cost |
|---|---|---|---|---|
| GPT-6 Luna | $0.20 | 3 | $0.88 | 1 |
| Particle-1 | $0.16 | 1 | $1.40 | 2 |
| Gemini 3.5 Flash-Lite | $0.85 | 8 | $1.44 | 3 ▲ cheaper than its price suggests |
| Claude Haiku 5.5 | $0.20 | 2 | $1.85 | 4 |
| MiniMax M3.1 Flash | $0.26 | 5 | $1.98 | 5 |
| Muse Ember | $0.20 | 4 | $2.13 | 6 |
| DeepSeek-V4.1-Flash-1017 | $0.52 | 6 | $3.73 | 7 |
| GPT-6.1 Terra | $2.00 | 10 | $5.69 | 8 |
| Claude Haiku 4.5 | $2.00 | 9 | $6.35 | 9 |
| GLM-5.3 FlashX Legal | $0.59 | 7 | $6.93 | 10 ▼ dearer than its price suggests |
| Mistral Medium 3.5 | $3.00 | 15 | $7.30 | 11 ▲ cheaper than its price suggests |
| Mistral Large 4 | $2.06 | 11 | $9.24 | 12 |
| Gemini 3.8 Flash | $3.00 | 13 | $9.43 | 13 |
| GPT-6.2 Sol | $4.00 | 18 | $12.45 | 14 ▲ cheaper than its price suggests |
| Gemini 4 Argon | $4.00 | 19 | $13.10 | 15 ▲ cheaper than its price suggests |
| GPT-6.1 Sol | $4.00 | 17 | $14.29 | 16 |
| Qwen4-Max | $2.25 | 12 | $20.37 | 17 ▼ dearer than its price suggests |
| Gemini 4 Carbon | $8.00 | 23 | $30.92 | 18 ▲ cheaper than its price suggests |
| Gemini 3.1 Pro Preview | $4.50 | 20 | $33.32 | 19 |
| Gemini 4 Hydrogen | $3.00 | 14 | $33.65 | 20 ▼ dearer than its price suggests |
| Muse Wonder | $8.00 | 24 | $35.42 | 21 ▲ cheaper than its price suggests |
| GPT-6.2 Astra | $16.00 | 26 | $36.60 | 22 ▲ cheaper than its price suggests |
| Claude Opus 5.5 | $8.00 | 22 | $45.71 | 23 |
| Kimi K3.1 Pro | $3.81 | 16 | $74.21 | 24 ▼ dearer than its price suggests |
| Gemini 4 Neon | $6.00 | 21 | $79.49 | 25 ▼ dearer than its price suggests |
| Claude Fable 5.2 | $16.00 | 25 | $151.05 | 26 |
ROASBench: a simulated year of paid-marketing plans (35 models)
| Model | Blended price | Rank by price | Measured cost per simulated year | Rank by cost |
|---|---|---|---|---|
| MiniMax M3.1 Flash | $0.26 | 5 | $0.00 | 1 ▲ cheaper than its price suggests |
| GPT-6 Luna | $0.20 | 3 | $0.01 | 2 |
| Particle-1 | $0.16 | 1 | $0.02 | 3 |
| Muse Ember | $0.20 | 4 | $0.03 | 4 |
| Claude Haiku 4.5 | $2.00 | 10 | $0.03 | 5 ▲ cheaper than its price suggests |
| Claude Haiku 5.5 | $0.20 | 2 | $0.03 | 6 ▼ dearer than its price suggests |
| Gemini 3.5 Flash-Lite | $0.85 | 8 | $0.04 | 7 |
| Mistral Large 4 | $2.06 | 13 | $0.07 | 8 ▲ cheaper than its price suggests |
| DeepSeek-V4.1-Flash-1017 | $0.52 | 6 | $0.08 | 9 ▼ dearer than its price suggests |
| GPT-6.1 Terra | $2.00 | 11 | $0.11 | 10 |
| GLM-5.3 | $2.15 | 14 | $0.13 | 11 ▲ cheaper than its price suggests |
| Mistral Medium 3.5 | $3.00 | 19 | $0.15 | 12 ▲ cheaper than its price suggests |
| GLM-5.3 FlashX Legal | $0.59 | 7 | $0.16 | 13 ▼ dearer than its price suggests |
| Gemini 3.8 Flash | $3.00 | 16 | $0.16 | 14 |
| Muse Spark 1.3 | $2.00 | 12 | $0.18 | 15 ▼ dearer than its price suggests |
| DeepSeek-V4-Pro (0423) | $1.77 | 9 | $0.20 | 16 ▼ dearer than its price suggests |
| GPT-6.2 Sol | $4.00 | 24 | $0.23 | 17 ▲ cheaper than its price suggests |
| Gemini 4 Argon | $4.00 | 25 | $0.24 | 18 ▲ cheaper than its price suggests |
| GPT-6 Sol | $4.00 | 23 | $0.24 | 19 ▲ cheaper than its price suggests |
| Claude Sonnet 5.5 | $4.00 | 22 | $0.42 | 20 |
| Grok 4.7 | $3.00 | 18 | $0.42 | 21 ▼ dearer than its price suggests |
| Gemini 4 Hydrogen | $3.00 | 17 | $0.44 | 22 ▼ dearer than its price suggests |
| Gemini 3.1 Pro Preview | $4.50 | 26 | $0.52 | 23 ▲ cheaper than its price suggests |
| Gemini 4 Carbon | $8.00 | 30 | $0.52 | 24 ▲ cheaper than its price suggests |
| Muse Wonder | $8.00 | 31 | $0.57 | 25 ▲ cheaper than its price suggests |
| Qwen4-Max | $2.25 | 15 | $0.69 | 26 ▼ dearer than its price suggests |
| Qwen3.8-Max (0902) | $3.00 | 20 | $0.73 | 27 ▼ dearer than its price suggests |
| GPT-6.2 Astra | $16.00 | 33 | $0.77 | 28 ▲ cheaper than its price suggests |
| Kimi K3 | $6.00 | 28 | $1.01 | 29 |
| Claude Opus 5.5 | $8.00 | 29 | $1.06 | 30 |
| GPT-6 Astra | $20.00 | 35 | $1.23 | 31 ▲ cheaper than its price suggests |
| Gemini 4 Neon | $6.00 | 27 | $1.28 | 32 ▼ dearer than its price suggests |
| Kimi K3.1 Pro | $3.81 | 21 | $1.60 | 33 ▼ dearer than its price suggests |
| Claude Fable 5.2 | $16.00 | 32 | $1.76 | 34 |
| Claude Fable 5.1 | $20.00 | 34 | $2.49 | 35 |
BulletBench: playing fast chess on the Lightning 10+1 clock (18 models)
| Model | Blended price | Rank by price | Measured cost per 1,000 games | Rank by cost |
|---|---|---|---|---|
| Seed-2.0-Mini | $0.18 | 1 | $1.93 | 1 |
| GPT-5.4 nano | $0.46 | 3 | $2.49 | 2 |
| GLM-5.3 FlashX Legal | $0.59 | 5 | $3.52 | 3 |
| Gemini 3.8 Flash | $3.00 | 13 | $4.05 | 4 ▲ cheaper than its price suggests |
| Mistral Small 4 | $0.26 | 2 | $4.09 | 5 ▼ dearer than its price suggests |
| Gemini 3.6 Flash | $3.00 | 12 | $5.00 | 6 ▲ cheaper than its price suggests |
| Gemini 3.1 Flash-Lite | $0.56 | 4 | $5.83 | 7 ▼ dearer than its price suggests |
| GPT-6.1 Terra | $2.00 | 10 | $6.42 | 8 |
| Gemini 3.5 Flash-Lite | $0.85 | 7 | $6.98 | 9 |
| Inkling Small | $0.64 | 6 | $7.68 | 10 ▼ dearer than its price suggests |
| Claude Sonnet 5.5 | $4.00 | 17 | $8.04 | 11 ▲ cheaper than its price suggests |
| Claude Haiku 4.5 | $2.00 | 9 | $9.53 | 12 ▼ dearer than its price suggests |
| Mistral Large 4 | $2.06 | 11 | $11.66 | 13 |
| GPT-5.4 mini | $1.69 | 8 | $13.62 | 14 ▼ dearer than its price suggests |
| Gemini 3.5 Flash | $3.38 | 16 | $21.30 | 15 |
| Gemini 4 Hydrogen | $3.00 | 14 | $21.95 | 16 |
| Mistral Medium 3.5 | $3.00 | 15 | $29.71 | 17 |
| GPT-6.1 Sol | $4.00 | 18 | $158.26 | 18 |
How to cut your LLM bill
Choose by cost per task, not price per token
Verbose models cost more than their price suggests. Run your own task on two or three candidates and compare what each run cost, as in the table above.
Use the smallest model that passes
Define what "good enough" means for your task, then test downwards. The cheapest model in each intelligence band is often within a few points of the best.
Set the reasoning effort deliberately
High reasoning can multiply the tokens a model writes. On our benchmarks it sometimes helps and sometimes makes results worse: Mistral Large 4 cost almost four times as much per listing at high reasoning, with fewer invented claims but more failures.
Cache the prompt you repeat
Most developers charge far less for cached input. Put instructions, examples and reference data at the start of the prompt, unchanged, so they can be cached.
Batch what can wait
Batch APIs from OpenAI, Anthropic, Google and others take about half off for work returned within a day: catalogue enrichment, evaluations, backfills.
Route by difficulty
Send easy requests to a cheap model and escalate only what it can't handle. A simple check on the cheap model's output often decides.
Cap the output
Ask for structured output and set a maximum length. Output tokens typically cost three to six times as much as input.
Watch launch discounts end
New models often launch at half price for a few weeks. Budget at the list price.
Quick answers
What is the cheapest LLM API?
Among models with an Artificial Analysis Intelligence Index score, Gemma 3 4B is the cheapest, at $0.042 per million tokens (three input tokens to one output). Cheaper models exist, but they are much less capable.
What is the cheapest capable LLM?
Claude Haiku 5.5 is the cheapest model scoring 40 or more on the Intelligence Index, at $0.20 per million tokens.
How much does the most capable model cost?
Claude Opus 5.5 scores highest on the Intelligence Index (57.6) and costs $4.00 per million input tokens and $20.00 per million output tokens.
Why do output tokens cost more than input tokens?
Generating each output token takes a full pass through the model, while input tokens are processed together. Across the models we track, output costs a median of four times as much as input.
What does 'blended price' mean?
One price per million tokens for a typical mix of three input tokens to one output, so models can be compared on a single number. Your own mix may differ: chat and summarisation are input-heavy, generation is output-heavy.
Which model is cheapest for your task?
Run the models you're considering side by side on your own prompt, with the cost of every answer. Or ask us to find the cheapest model that's good enough on your real data.
Sources and method
- Developer prices: read by hand from each developer's own pricing page, checked 7 Oct 2026; each links from its row in the table. Launch discounts and their end dates are noted.
- Host prices: every host's price for each model from OpenRouter's public endpoints list, checked 10 Nov 2026. "Typical host" is the median; "cheapest host" is the lowest blended price, with the model's quantisation where the host states it.
- Intelligence and speed: Artificial Analysis (data sourced from Artificial Analysis). Our overall: the Spring Prompt overall score.
- Measured cost per task: what our own benchmark runs cost (CatalogBench, DeckBench, SurveyBench, ROASBench, BulletBench), from provider billing, at each model's cheapest setting.